Livestock grazing can be a problem for forest conservation because it can generate heterogeneous and unpredictable changes in plant communities. Understanding these changes is important for generating management strategies that are compatible with long-term conservation of threatened forests. Livestock exclusion is a useful experimental approach used to evaluate grazing effects. However, the evidence showing the effects of grazing on forests is mixed and little in know about the responses of different plant life forms, especially in dry forests. In this study, we evaluated the effects of a 7-8 year of livestock exclusion experiment on understory plant community structure in the dry Chaco forest (Argentina). We categorized understory plant life forms as shrubs, succulents (Cactaceae family + Bromelia hieronymi), and herbs (forbs, grasses and vines). Then, we compared the plant community structure (richness, diversity, density and cover) and understory structure (soil hardness, bare soil and vegetation vertical and horizontal structure) between five excluded plots and five grazed plots, in a paired design. We found that livestock exclusion lead to an increase in grass species richness and grass cover as well as an increase in lower understory biomass (0-0.5 m) and a decrease in percentage of bare soil. On excluded plots, dominant herbs were Setaria nicorae (grass), Trichloris crinita (grass), and Justicia squarrosa (forb). Grass species that were recorded exclusively on excluded plots were Gouinia latifolia, T. crinita, and Pappophorum mucronulatum, all forage species preferred by livestock. In contrast, on grazed plots, the dominant species was Stenandrium dulce (forb), a species with resistance strategies to grazing. As for the other variables, we did not find strong differences between excluded and grazed plots. Livestock grazing did not modify the ensemble structure of shrubs and succulents nor did it change the horizontal vegetation structure or soil hardness. Our evidence suggests that the assemblage composed by shrubs and succulents seems to be tolerant to livestock grazing, and that the grass assemblage has the ability to quickly recover when grazing stops. Finally, the effectiveness of exclusion as a management tool will depend on which attribute of the plant community to be conserved or recovered. In dry Chaco forests after many years of grazing at moderate stocking rates, livestock exclusion could help recover grass cover, generate opportunities for the establishment of certain grass species that are sensitive to grazing, and increase ground cover.